Summit 2011 Working Sessions/Session001

From OWASP
Revision as of 13:34, 16 January 2011 by Sarah Baso (Talk | contribs)

Jump to: navigation, search

T. browser security.jpg

Google-groups-logo-1.jpgJoin the Google Group for this track

About

The Browser Security track of the OWASP Summit 2011 was an initial community effort to bring together browser vendors, major web app providers, well-known white hat hackers, and OWASP leaders to discuss what can be done to enhance web security through the browser. The track comprised of two half-day day workshops on chosen subtopics (see tabs). We invited some of the world's top experts to maximize the chances of moving forward this important area or application security.

Session Notes

Here are the notes from all the four browser security sessions. Later on we will publish a Browser Security Report building on these sessions.

Site Security Policy notes (pdf)
DOM Sandboxing notes (pdf)
HTML5 Security notes (pdf)
EcmaScript 5 Security notes (pdf)
Enduser Warnings notes (pdf)

Who Participated?

The following browser, major web app, and plugin vendors participated in the browser security sessions:
Chrome_small.jpg Firefox_small.jpg Internet_explorer_small.jpg Paypal_logo.gif Adobe_logo_standard_for_Tasha.jpg

/John Wilander, Session Chair

DOM Sandboxing

Virtualization and Sandboxing for Secure Multi-Domain Web Apps

JS DOM Box Jasvir Gaz.jpg

Co-chair Dr Jasvir Nagra

Jasvir Nagra is a researcher and software engineer at Google. He is the designer of Caja - a secure subset of HTML, CSS and JavaScript; co-author of Surreptitious Software - a book on obfuscation, software watermarking and tamper-proofing, contributer to Shindig - the reference implementation of OpenSocial.

Co-chair Gareth Heyes

Gareth "Gaz" Heyes calls himself Chief Conspiracy theorist and is affiliated with Microsoft. He is the designer and developer behind JSReg – a Javascript sandbox which converts code using regular expressions; HTMLReg & CSSReg – converters of malicious HTML/CSS into a safe form of HTML. He is also one of the co-authors of Web Application Obfuscation: '-/WAFs..Evasion..Filters//alert(/Obfuscation/)-' – a book on how an attacker would bypass different types of security controls including IDS/IPS.

Subjects and Goals (draft)

Goals and issues that need browser vendor cooperation:

  • Attenuated versions of existing apis to sandboxed code. How should browsers introduce new apis into the sandbox or allow the sandbox to provide attenuated versions of existing apis to sandboxed code? For example, lets say the sandbox wants to provide an attenuated "alert" function to sandboxed code which does something slightly different than the real "alert". What kind of apis could the browser provide to safely allow such extensions/apis? Do these need to be standardized such that different sandbox vendors can interoperate.
  • Client side sandboxed apps maintaining state and authentication. For example if a user is created in a sandboxed app how is it determined what that user can do?
  • Create a standard for modifying a sandboxed environment
  • Deprecate and discourage standards which ambiently or undeniably pass credentials.
  • Adopt a simpler rights amplification api like Web Introducer
  • Create a standard for authentication within a sandboxed environment (maybe interfacing with existing auth without passing creds like 0Auth works)

Working Form

The working form will most probably be short presentations to frame the topic and then round table discussions. Depending on number of attendees we'll break into groups.


HTML5

HTML5 Security

Html5 mario hackvertor.jpg

Co-chair Mario Heiderich

Mario Heiderich works as a researcher for the Ruhr-University in Bochum, Germany and currently focuses on HTML5, SVG security and security implications of the ES5 specification draft. Mario invoked the HTML5 security cheat-sheet and maintains the PHPIDS filter rules. In his spare time he delivers trainings and security consultancy for larger German and international companies. He is also one of the co-authors of Web Application Obfuscation: '-/WAFs..Evasion..Filters//alert(/Obfuscation/)-' – a book on how an attacker would bypass different types of security controls including IDS/IPS.

Co-chair Gareth Heyes

Gareth "Gaz" Heyes calls himself Chief Conspiracy theorist and is affiliated with Microsoft. He is the designer and developer behind JSReg – a Javascript sandbox which converts code using regular expressions; HTMLReg & CSSReg – converters of malicious HTML/CSS into a safe form of HTML. He is also one of the co-authors of Web Application Obfuscation: '-/WAFs..Evasion..Filters//alert(/Obfuscation/)-' – a book on how an attacker would bypass different types of security controls including IDS/IPS.

Subjects and Goals (draft)

  • Handle autofocus in a unified and secure way. Make sure SOP applies for autofocus usage in frame/iframe'd websites. Re-discuss necessity for (future) attributes like this.
  • Discuss necessity and capability for the HTML5 form controls. Do we need a non-SOP formaction attribute and why?
  • Goal I: Initiate and create documentation and references for developers that address security issues. Html5sec.org is a start but impossible to continue or extend large scale without vendor help
  • Goal II: Discuss and heavily restrict SVG capabilities - especially when deployed in CSS backgrounds and <img> tags. Mainly Opera and Mozilla are addressed here.
  • Long Term Goal(s): Provide a working and easy to use as well as vendor supported HTML5 compliant filter software such as HTMLPurifier. Browser vendors should participate in creating security software and filters - not undermine them as we could experience in the last decade

EcmaScript 5

EcmaScript 5 Security

Co-chair Mario Heiderich

Mario Heiderich works as a researcher for the Ruhr-University in Bochum, Germany and currently focuses on HTML5, SVG security and security implications of the ES5 specification draft. Mario invoked the HTML5 security cheat-sheet and maintains the PHPIDS filter rules. In his spare time he delivers trainings and security consultancy for larger German and international companies. He is also one of the co-authors of Web Application Obfuscation: '-/WAFs..Evasion..Filters//alert(/Obfuscation/)-' – a book on how an attacker would bypass different types of security controls including IDS/IPS.

Co-chair 2

To be confirmed.

Subjects and Goals (draft)

  • Fix the problems with Object.defineProperty() and property unsealing / double-freezing. Implement it if not yet done.
  • Goal I: Raise awareness for the power or object freezing in a security context. ES5 can really make a change here.
  • Goal II: Raise awareness in seeing the DOM as the place where XSS attacks actually take place - and where they should be prevented. CSP is a great yet still immature start - but worth discussing and extending. Discuss specification drafts for a secure DOM and easy to configure capability profiles with reasonable and quantitative proofs of concept.
  • Long Term Goal: Discuss the possibility of vendor supported client side security mechanisms. Client side IDS/IPS based on ES5 can be possible - yet have to be designed and specified.


Site Security Policy

There are several initiatives for expressing and enforcing website security policies. HTTP Strict Transport Security for enforced TLS. Content Security Policy for whitelisting resource domains and enforcing file-only JavaScript. X-Frame-Options for enforce framing restrictions. Harmonizing these features among browsers is a huge task. Getting developers to adopt and implement is even more challenging. This session will try to address all of these questions as well as the technical alternatives – headers, meta tags, nonces, signatures, html zones, css-like policies, violation events etc.

  • Should we have independent, coherent and simple policy mechanisms or a generalized, extensible policy mechanism?
  • Should developers have multiple choices for expressing policies such as headers and meta tags?
  • Should policies restrict domains, URLs, or elements? What are the consequences?
  • Should one or two browser vendors deploy a policy mechanism in the field, collect experience, and then we set a standard?
  • How do we help developers understand the need for policies and how do we help them write/generate/maintain policies?
  • How important is performance and web 1.0/web 2.0 compliance? How much of the web can we afford to break? 0 %?

Co-chair Jeff Hodges

Jeff Hodges is Distinguished Security Engineer at PayPal, Inc and one of the three original authors of the HTTP Strict Transport Security spec. Check out IdentityMeme.org Jeff's blog.

Co-chair David Ross

David Ross is a Principal Security Software Engineer on the MSRC Engineering team at Microsoft. Prior to joining MSRC Engineering in 2002, David spent his formative years on the Internet Explorer Security Team and wears the battle scars with pride. Check out David’s blog.

Co-chair Lucas Adamski

Lucas Adamski is the Director of Security Engineering at Mozilla Corporation. He specializes in penetration testing, incident response, spec and code reviews, fuzzing, training, security processes and software PLC. Check out Lucas' blog.

Enduser Warnings

Enduser Warnings

Three browsers user info.jpg

Subjects and Goals (draft)

Clearly there is a need for warnings that users understand and that conveys the right information. Perhaps we can agree on some guidelines or at least exchange lessons learned.

  • How should browsers signal invalid SSL certs to the enduser? Are we helping security right now? What to do about 50 % of users clicking through warnings? Mozilla replaces the padlock with a site identity button i Firefox 4. "Larry" will inform the user of the site's status. Google recently tried out a skull & bones icon for bad certs but moved back to padlocks again.
  • How should browsers communicate other kinds of information such as privacy, malware warnings, "not visited before" etc? Forbes had an interesting example of how to visualize privacy.

Some additional information, thoughts and discussions on these subjects elsewhere:


Securing Plugins

Securing Plugins

Should browsers ship with default plugins? Should plugins be auto-updated? Can plugins or versions of plugins be blacklisted centrally?


Blacklisting

Blacklisting

Can we cooperate better on blacklisting? Does it work between cultures, i e can we have the same process for reporting throughout the world?


OS Integration

OS Integration

More and more features in browsers get integrated with the underlying operating system. Processes, fonts, filesystem, 3D graphics. How do we secure this?

Sandboxed Browser

Sandboxed Tabs/Domains/Browser

Microsoft Research has been doing some groundbreaking work on the Gazelle browser, Chrome uses a sandboxing model, and the IronSuite provides sandboxed versions of Firefox (IronFox) and Safari on Mac OS X.


Questions? Contact John Wilander, Session Chair

Return Global Summit 2011 Home Page
Return to Global Summit 2011 Working Sessions